cp-pc / Distributional-Discrepancy

Distributional Discrepancy: A Metric for Unconditional Text Generation

Code for: Distributional Discrepancy: A Metric for Unconditional Text Generation

Requirement

We suggest you run the platform under Python 3.6+ with following libs:

• tensorflow>=1.12.0
• tensorflow_hub>=0.7.0
• numpy 1.16.4
• scipy 1.3.1
• nltk 3.4.5
• colorama 0.4.1
• CUDA 7.5+ (Suggested for GPU speed up, not compulsory)

Or just type pip install -r requirements.txt in your terminal.

Get Started

git clone https://github.com/anonymous1100/Distributional-Discrepancy.git
cd Distributional-Discrepancy
# run with default setting
python main.py

#You can also change the model and data in main.py and then run main.py

Structure

• data folder: emnlp_news and image_coco dataset
• models folder: Training codes for LSTM and GPT-2
• utils folder: code for some evaluation
• experiments folder: Experimental results
  • experiments/experiment_name/tmp folder: Generated samples
  • experiments/experiment_name/output folder: some output files
  • experiments/experiment_name/ckpts folder: Trained model
  • experiments/experiment_name/summary folder: tensorboard record

Contact

For any questions, feel free to open an issue via github, or to send me an email at
[e-mail].